accelerate the molecular docking process with slurm
Published on: 2025-04-26T08:40:43
Table Of Contents
概览
尝试将一个分子对接流程并行化
zhihao@dust|/home/zhihao/Downloads/mol-docking/Test|$ ls
2w17.pdb conf.txt pro.pdb receptor.pdbqt testset200.sdf vina.sh
zhihao@dust|/home/zhihao/Downloads/mol-docking/Test|$集群搭建使用 virtualbox, 镜像为 Rocky-8.10-x86_64-dvd1.iso。共构建 3 节点，包含一个控制节点（ 主机名 slurm-controller ），两个计算节点（ 主机名 slurm-compute[1,2] ），配置均为 4 cpu, 3G mem, 50G ssd。slurm database 服务部署在控制节点上。集群使用 /mnt/slurm_shared/ nfs 共享文件夹。
ansible 自动部署 slurm 集群
集群安装ansible
zhihao@slurm-controller|/home/zhihao/.ssh|$ sudo yum install ansible -y
[sudo] password for zhihao:
Last metadata expiration check: 1:15:45 ago on Sat 26 Apr 2025 04:30:06 AM EDT.
Dependencies resolved.
=================================================================================================================================
Package Architecture Version Repository Size
=================================================================================================================================
Installing:
ansible noarch 9.2.0-1.el8 epel 46 M
Installing dependencies:
ansible-core x86_64 2.16.3-2.el8 appstream 3.6 M
git-core x86_64 2.43.5-2.el8_10 appstream 11 M
mpdecimal x86_64 2.5.1-3.el8 appstream 92 k
python3.12 x86_64 3.12.8-1.el8_10 appstream 29 k
python3.12-cffi x86_64 1.16.0-2.el8 appstream 298 k
python3.12-cryptography x86_64 41.0.7-1.el8 appstream 1.2 M
python3.12-libs x86_64 3.12.8-1.el8_10 appstream 10 M
python3.12-pip-wheel noarch 23.2.1-4.el8 appstream 1.5 M
python3.12-ply noarch 3.11-2.el8 appstream 133 k
python3.12-pycparser noarch 2.20-2.el8 appstream 144 k
python3.12-pyyaml x86_64 6.0.1-2.el8 appstream 202 k
sshpass x86_64 1.09-4.el8 appstream 29 k
Installing weak dependencies:
python3-jmespath noarch 0.9.0-11.el8 appstream 44 k
Transaction Summary
=================================================================================================================================
Install 14 Packages
Total download size: 75 M
Installed size: 560 M
Downloading Packages:
...
zhihao@slurm-controller|/home/zhihao/.ssh|$ ssh -t slurm-compute1 sudo yum install ansible -y
...
zhihao@slurm-controller|/home/zhihao/.ssh|$ ssh -t slurm-compute2 sudo yum install ansible -y
...创建 ansible 目录，创建主机清单
zhihao@slurm-controller|/home/zhihao/Downloads|$ mkdir -p slurm-cluster
zhihao@slurm-controller|/home/zhihao/Downloads|$ cd slurm-cluster/
/home/zhihao/Downloads/slurm-cluster
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$ cat inventory.ini
[slurm_controllers]
slurm-controller ansible_host=192.168.1.100
[slurm_database]
slurm-controller ansible_host=192.168.1.100
[slurm_compute_nodes]
slurm-compute1 ansible_host=192.168.1.101
slurm-compute2 ansible_host=192.168.1.102
[slurm:children]
slurm_controllers
slurm_compute_nodes
[all:vars]
ansible_user=root
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$创建集群 playbook
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$ cat slurm-cluster.yml
- name: configure common roles
hosts: slurm
become: yes
roles:
- common
- name: configure slurm controller
hosts: slurm_controllers
become: yes
roles:
- slurm-controller
- name: configure slurm compute nodes
hosts: slurm_compute_nodes
become: yes
roles:
- slurm-compute-node
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$创建角色目录
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$ mkdir -p roles/common/{tasks,templates}
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$ mkdir -p roles/slurm-controller/{tasks,templates}
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$ mkdir -p roles/slurm-compute-node/{tasks,templates}
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$编写 common 角色，添加 epel、powertools 仓库，创建 slurm 用户
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$ cat roles/common/tasks/main.yml
- name: install epel repository
dnf:
name: epel-release
state: present
- name: enable powertools repository
ansible.builtin.command: >
dnf config-manager --set-enabled powertools
- name: create slurm group
group:
name: "slurm"
gid: "1001"
- name: create slurm user
user:
name: "slurm"
uid: "1001"
group: "slurm"
shell: /bin/false
home: /nonexistent
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$编写 slurm-controller 角色
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$ cat roles/slurm-controller/tasks/main.yml
- name: install slurm packages
dnf:
name:
- slurm
- slurm-slurmd
- slurm-slurmdbd
- slurm-slurmctld
- mariadb
- mariadb-server
- python3-mysqlclient
- munge
state: present
- name: change owner and group of slurm dirs
file:
path: "{{ item }}"
owner: slurm
group: slurm
recurse: yes
loop:
- /var/log/slurm
- /var/spool/slurm/d
- /var/spool/slurm/ctld
- name: generate munge key
command: create-munge-key
args:
creates: /etc/munge/munge.key
- name: start and enable services
systemd:
name: "{{ item }}"
state: started
enabled: yes
loop:
- munge
- mariadb
- name: create slurm accounting database
mysql_db:
name: "slurm_acct_db"
state: present
become: true
- name: create slurm database user and grant privileges
mysql_user:
name: "slurm"
password: "123456"
priv: "slurm_acct_db.*:ALL"
host: "slurm-controller"
state: present
become: true
- name: configure slurmdbd.conf
template:
src: slurmdbd.conf.j2
dest: /etc/slurm/slurmdbd.conf
owner: slurm
group: slurm
mode: '0600'
backup: yes
- name: start and enable slurmdbd service
systemd:
name: slurmdbd
state: restarted
enabled: yes
become: true
- name: configure slurm
template:
src: slurm.conf.j2
dest: /etc/slurm/slurm.conf
owner: slurm
group: slurm
mode: '0644'
backup: yes
- name: start and enable slurmdbd service
systemd:
name: slurmctld
state: restarted
enabled: yes
become: true
- name: configure firewall
firewalld:
port: "{{ item }}"
state: enabled
permanent: yes
immediate: yes
loop:
- 6817/tcp
- 32768-60999/tcp
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$编写 slurm-compute-node 角色
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$ cat roles/slurm-compute-node/tasks/main.yml
- name: copy munge key from slurm-controller
become: yes
synchronize:
src: /etc/munge/munge.key
dest: /etc/munge/munge.key
delegate_to: slurm-controller
- name: install slurm packages
dnf:
name:
- slurm
- slurm-slurmd
- munge
state: present
- name: copy slurm from slurm controller
synchronize:
src: /etc/slurm/slurm.conf
dest: /etc/slurm/slurm.conf
delegate_to: slurm-controller
- name: start and enable services
systemd:
name: "{{ item }}"
state: restarted
enabled: yes
loop:
- slurmd
- munge
- name: change owner and group of slurm dirs
file:
path: "{{ item }}"
owner: slurm
group: slurm
recurse: yes
loop:
- /var/log/slurm
- /var/spool/slurm/d
- /var/spool/slurm/ctld
- name: configure firewall
firewalld:
port: "{{ item }}"
state: enabled
permanent: yes
immediate: yes
loop:
- 6818/tcp
- 32768-60999/tcp
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$创建 slurm controller 模板
[zhihao@slurm-controller slurm-cluster]$ cat roles/slurm-controller/templates/slurm.conf.j2
ControlMachine={{ inventory_hostname }}
AuthType=auth/munge
CryptoType=crypto/munge
MpiDefault=pmix
ProctrackType=proctrack/cgroup
ReturnToService=1
SlurmctldPidFile=/var/run/slurm/slurmctld.pid
SlurmctldPort=6817
SlurmdPidFile=/var/run/slurm/slurmd.pid
SlurmdPort=6818
SlurmdSpoolDir=/var/spool/slurm/d
SlurmUser=slurm
StateSaveLocation=/var/spool/slurm/ctld
SwitchType=switch/none
TaskPlugin=task/none
InactiveLimit=0
KillWait=30
MinJobAge=300
SlurmctldTimeout=120
SlurmdTimeout=300
Waittime=0
SchedulerType=sched/backfill
SelectType=select/cons_res
SelectTypeParameters=CR_CPU
AccountingStorageEnforce=limits
AccountingStorageHost={{ inventory_hostname }}
AccountingStorageType=accounting_storage/slurmdbd
AccountingStorageUser=slurm
AccountingStoreJobComment=YES
ClusterName=rocky_cluster
JobCompType=jobcomp/none
JobAcctGatherFrequency=30
JobAcctGatherType=jobacct_gather/none
SlurmctldDebug=3
SlurmctldLogFile=/var/log/slurm/slurmctld.log
SlurmdDebug=3
SlurmdLogFile=/var/log/slurm/slurmd.log
# Nodes Configuration
NodeName=slurm-compute1 CPUs=4 Boards=1 SocketsPerBoard=1 CoresPerSocket=4 ThreadsPerCore=1 RealMemory=2786
NodeName=slurm-compute2 CPUs=4 Boards=1 SocketsPerBoard=1 CoresPerSocket=4 ThreadsPerCore=1 RealMemory=2786
PartitionName=debug Nodes=slurm-compute[1-2] Default=YES MaxTime=INFINITE State=UP
[zhihao@slurm-controller slurm-cluster]$ 创建 slurm database 模板
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$ cat roles/slurm-controller/templates/slurmdbd.conf.j2
AuthType=auth/munge
DebugLevel=4
DbdHost={{ inventory_hostname }}
LogFile=/var/log/slurm/slurmdbd.log
PidFile=/var/run/slurm/slurmdbd.pid
PurgeEventAfter=1month
PurgeJobAfter=1month
PurgeResvAfter=1month
PurgeStepAfter=1month
PurgeSuspendAfter=1month
PurgeTXNAfter=1month
PurgeUsageAfter=1month
SlurmUser=slurm
StorageType=accounting_storage/mysql
StorageHost={{ inventory_hostname }}
StoragePass=123456
StorageUser=slurm
StorageLoc=slurm_acct_db
Parameters=PreserveCaseUser
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$运行 slurm playbook
[root@slurm-controller slurm-cluster]# ansible-playbook -i inventory.ini slurm-cluster.yml
PLAY [configure common roles] ********************************************************
TASK [Gathering Facts] ***************************************************************
ok: [slurm-compute1]
ok: [slurm-compute2]
ok: [slurm-controller]
TASK [common : install epel repository] **********************************************
ok: [slurm-controller]
ok: [slurm-compute2]
ok: [slurm-compute1]
TASK [common : enable powertools repository] *****************************************
changed: [slurm-compute1]
changed: [slurm-controller]
changed: [slurm-compute2]
TASK [common : create slurm group] ***************************************************
ok: [slurm-compute1]
ok: [slurm-controller]
ok: [slurm-compute2]
TASK [common : create slurm user] ****************************************************
ok: [slurm-compute1]
ok: [slurm-controller]
ok: [slurm-compute2]
PLAY [configure slurm controller] ****************************************************
TASK [Gathering Facts] ***************************************************************
ok: [slurm-controller]
TASK [slurm-controller : install slurm packages] *************************************
ok: [slurm-controller]
TASK [slurm-controller : change owner and group of slurm dirs] ***********************
ok: [slurm-controller] => (item=/var/log/slurm)
ok: [slurm-controller] => (item=/var/spool/slurm/d)
ok: [slurm-controller] => (item=/var/spool/slurm/ctld)
TASK [slurm-controller : generate munge key] *****************************************
ok: [slurm-controller]
TASK [slurm-controller : start and enable services] **********************************
ok: [slurm-controller] => (item=munge)
ok: [slurm-controller] => (item=mariadb)
TASK [slurm-controller : create slurm accounting database] ***************************
ok: [slurm-controller]
TASK [slurm-controller : create slurm database user and grant privileges] ************
[WARNING]: Option column_case_sensitive is not provided. The default is now false, so
the column's name will be uppercased. The default will be changed to true in
community.mysql 4.0.0.
ok: [slurm-controller]
TASK [slurm-controller : configure slurmdbd.conf] ************************************
ok: [slurm-controller]
TASK [slurm-controller : start and enable slurmdbd service] **************************
changed: [slurm-controller]
TASK [slurm-controller : configure slurm] ********************************************
ok: [slurm-controller]
TASK [slurm-controller : start and enable slurmdbd service] **************************
changed: [slurm-controller]
TASK [slurm-controller : configure firewall] *****************************************
ok: [slurm-controller] => (item=6817/tcp)
ok: [slurm-controller] => (item=32768-60999/tcp)
PLAY [configure slurm compute nodes] *************************************************
TASK [Gathering Facts] ***************************************************************
ok: [slurm-compute1]
ok: [slurm-compute2]
TASK [slurm-compute-node : copy munge key from slurm-controller] *********************
ok: [slurm-compute1 -> slurm-controller(192.168.1.100)]
ok: [slurm-compute2 -> slurm-controller(192.168.1.100)]
TASK [slurm-compute-node : install slurm packages] ***********************************
ok: [slurm-compute1]
ok: [slurm-compute2]
TASK [slurm-compute-node : copy slurm from slurm controller] *************************
ok: [slurm-compute1 -> slurm-controller(192.168.1.100)]
ok: [slurm-compute2 -> slurm-controller(192.168.1.100)]
TASK [slurm-compute-node : start and enable services] ********************************
changed: [slurm-compute1] => (item=slurmd)
changed: [slurm-compute2] => (item=slurmd)
changed: [slurm-compute1] => (item=munge)
changed: [slurm-compute2] => (item=munge)
TASK [slurm-compute-node : change owner and group of slurm dirs] *********************
ok: [slurm-compute1] => (item=/var/log/slurm)
ok: [slurm-compute2] => (item=/var/log/slurm)
changed: [slurm-compute1] => (item=/var/spool/slurm/d)
changed: [slurm-compute2] => (item=/var/spool/slurm/d)
ok: [slurm-compute1] => (item=/var/spool/slurm/ctld)
ok: [slurm-compute2] => (item=/var/spool/slurm/ctld)
TASK [slurm-compute-node : configure firewall] ***************************************
ok: [slurm-compute1] => (item=6818/tcp)
ok: [slurm-compute2] => (item=6818/tcp)
ok: [slurm-compute1] => (item=32768-60999/tcp)
ok: [slurm-compute2] => (item=32768-60999/tcp)
PLAY RECAP ***************************************************************************
slurm-compute1 : ok=12 changed=3 unreachable=0 failed=0 skipped=0 rescued=0 ignored=0
slurm-compute2 : ok=12 changed=3 unreachable=0 failed=0 skipped=0 rescued=0 ignored=0
slurm-controller : ok=17 changed=3 unreachable=0 failed=0 skipped=0 rescued=0 ignored=0
[root@slurm-controller slurm-cluster]#检查集群状态
[root@slurm-controller slurm-cluster]# sinfo
PARTITION AVAIL TIMELIMIT NODES STATE NODELIST
debug* up infinite 2 idle slurm-compute[1-2]
[root@slurm-controller slurm-cluster]# sacct
JobID JobName Partition Account AllocCPUS State ExitCode
------------ ---------- ---------- ---------- ---------- ---------- --------
[root@slurm-controller slurm-cluster]# squeue
JOBID PARTITION NAME USER ST TIME NODES NODELIST(REASON)
[root@slurm-controller slurm-cluster]#
[root@slurm-controller log]# srun sleep 8
[root@slurm-controller slurm-cluster]# squeue
JOBID PARTITION NAME USER ST TIME NODES NODELIST(REASON)
44 debug sleep root R 0:03 1 slurm-compute1
[root@slurm-controller slurm-cluster]# squeue
JOBID PARTITION NAME USER ST TIME NODES NODELIST(REASON)
[root@slurm-controller slurm-cluster]#添加zhihao为管理员
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$ sudo sacctmgr add account name=zhihao
Adding Account(s)
zhihao
Settings
Description = Account Name
Organization = Parent/Account Name
Associations
A = zhihao C = rocky_clus
Would you like to commit changes? (You have 30 seconds to decide)
(N/y): y
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$ sudo sacctmgr add user name=zhihao account=zhihao AdminLevel=Admin
Adding User(s)
zhihao
Settings =
Admin Level = Administrator
Associations =
U = zhihao A = zhihao C = rocky_clus
Non Default Settings
Would you like to commit changes? (You have 30 seconds to decide)
(N/y): y
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$ sacctmgr show user User Def Acct Admin
---------- ---------- ---------
root root Administ+
zhihao zhihao Administ+
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-cluster|$
运行 vina demo
检查脚本内容
[root@slurm-controller Test]# cat vina.sh
# This is the serial script
export PATH=/home/user/Programs/autodock/autodock_vina_1_1_2_linux_x86/bin/:$PATH
export PATH=/home/user/Programs/autodock/MGLTools-1.5.6/MGLToolsPckgs/AutoDockTools/Utilities24/:$PATH
if [ -d "mols" ]; then
rm -rf mols
fi
mkdir mols
prepare_receptor4.py -r pro.pdb -e -o receptor.pdbqt ########## prepare only once
obabel -isd testset.sdf -omol2 -O testset.mol2
cd mols
obabel ../testset.mol2 -O vs.mol2 -m
for mol2 in `ls *.mol2` ############# loop by every small molecules
do
prepare_ligand4.py -l $mol2 -o ligand.pdbqt
vina --config ../conf.txt --cpu 1 --out out.pdbqt --log log.txt
obabel -ipdbqt out.pdbqt -osd -O temp.sdf
cat temp.sdf >> output.sdf
rm log.txt ligand.pdbqt out.pdbqt temp.sdf
done
cd ../
[root@slurm-controller Test]#脚本结构为：
- 受体蛋白文件为 pdb, 转换为 pdbqt 格式供 vina 使用
- 小分子结构文件 testset.sdf 转换成 testset.mol2 格式
- testset.mol2 多分子拆分为 vs<id>.mol2 单分子文件
- 循环
- mol2 转为 pdbqt
- vina 对配体、受体 pdbqt 文件对接，输出 pdbqt 对接结果
- pdbqt 对接结果转回 sdf
- 收集 sdf 结果文件
尝试运行脚本失败
[root@slurm-controller Test]# ./vina.sh
./vina.sh: line 8: prepare_receptor4.py: command not found
./vina.sh: line 9: obabel: command not found
./vina.sh: line 11: obabel: command not found
ls: cannot access '*.mol2': No such file or directory
[root@slurm-controller Test]#下载解压 autodock_vina_1_1_2_linux_x86.tgz, mgltools_x86_64Linux2_1.5.6.tar_.gz
[root@slurm-controller ~]# wget "https://vina.scripps.edu/wp-content/uploads/sites/55/2020/12/autodock_vina_1_1_2_linux_x86.tgz"
--2025-04-27 09:11:13-- https://vina.scripps.edu/wp-content/uploads/sites/55/2020/12/autodock_vina_1_1_2_linux_x86.tgz
Resolving vina.scripps.edu (vina.scripps.edu)... 192.26.252.19
Connecting to vina.scripps.edu (vina.scripps.edu)|192.26.252.19|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 1238242 (1.2M) [application/x-gzip]
Saving to: ‘autodock_vina_1_1_2_linux_x86.tgz’
autodock_vina_1_1_2_l 100%[=======================>] 1.18M 23.4KB/s in 51s
2025-04-27 09:12:05 (23.7 KB/s) - ‘autodock_vina_1_1_2_linux_x86.tgz’ saved [1238242/1238242]
zhihao@slurm-controller|/home/zhihao/Downloads|$ tar -zxvf autodock_vina_1_1_2_linux_x86.tgz
autodock_vina_1_1_2_linux_x86/
autodock_vina_1_1_2_linux_x86/LICENSE
autodock_vina_1_1_2_linux_x86/bin/
autodock_vina_1_1_2_linux_x86/bin/vina
autodock_vina_1_1_2_linux_x86/bin/vina_split
zhihao@slurm-controller|/home/zhihao/Downloads|$ ls auto
autodock_vina_1_1_2_linux_x86/ autodock_vina_1_1_2_linux_x86.tgz
zhihao@slurm-controller|/home/zhihao/Downloads|$ ls autodock_vina_1_1_2_linux_x86/
bin LICENSE
zhihao@slurm-controller|/home/zhihao/Downloads|$
zhihao@slurm-controller|/home/zhihao/Downloads|$ tar -zxf mgltools_x86_64Linux2_1.5.6.tar_.gz
zhihao@slurm-controller|/home/zhihao/Downloads|$ ls mgl
mgltools_x86_64Linux2_1.5.6/ mgltools_x86_64Linux2_1.5.6.tar_.gz*
zhihao@slurm-controller|/home/zhihao/Downloads|$ ls mgltools_x86_64Linux2_1.5.6
Data.tar.gz LICENSES Python2.5_x86_64Linux2.tar.gz Tools
install.sh MGLToolsPckgs.tar.gz README
zhihao@slurm-controller|/home/zhihao/Downloads|$安装 mgltools
zhihao@slurm-controller|/home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6|$ ./install.sh
...
tk8.4/safetk.tcl
tk8.4/mkpsenc.tcl
tk8.4/xmfbox.tcl
Python installed, please wait for the rest of MGLTools to be installed
Running /home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/bin/python Tools/install.py
Installing MGLPackages
Installing files from MGLToolsPckgs .tar.gz
eInstalling files from Data .tar.gz
Creating the pmv, adt, vision and tester scripts
current directory: /home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6
setting PYTHONHOME environment
Traceback (most recent call last):
File "/home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/MGLToolsPckgs/mglutil/splashregister/license.py", line 7, in <module>
tk_root = Tkinter.Tk()
File "/home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/lib/python2.5/lib-tk/Tkinter.py", line 1647, in __init__
self.tk = _tkinter.create(screenName, baseName, className, interactive, wantobjects, useTk, sync, use)
_tkinter.TclError: couldn't connect to display ":0"
MGLTools installation complete.
To run pmv, adt, vision or pythonsh scripts located at:
/home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/bin
you will need to do ONE of the following:
-- add the /home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/bin to the path environment variable in .cshrc or .bashrc:
.cshrc:
set path = (/home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/bin $path)
.bashrc
export PATH=/home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/bin:$PATH
-- create aliases in your .cshrc or .bashrc
.cshrc:
alias pmv /home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/bin/pmv
alias adt /home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/bin/adt
alias vision /home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/bin/vision
alias pythonsh /home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/bin/pythonsh
.bashrc
alias pmv='/home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/bin/pmv'
alias adt='/home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/bin/adt'
alias vision='/home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/bin/vision'
alias pythonsh='/home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/bin/pythonsh'
-- source ./initMGLtools.sh (bash) or ./initMGLtools.csh (c-shell)
Please have a look at README file for more information about
licenses, tutorials, documentations and mailing lists for the different
packages enclosed in this distribution
If you have any problems please visit our FAQ page (http://mgltools.scripps.edu/documentation/faq).
zhihao@slurm-controller|/home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6|$ 更新 vina.sh 环境变量
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-lab/Test|$ head -3 vina.sh
# This is the serial script
export PATH=/home/zhihao/Downloads/autodock_vina_1_1_2_linux_x86/bin:$PATH
export PATH=/home/zhihao/Downloads/mgltools_x86_64Linux2_1.5.6/bin:$PATH
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-lab/Test|$安装 openbabel-3.1.1-18.el8.x86_64
zhihao@slurm-controller|/home/zhihao/Downloads|$ sudo dnf install openbabel
...环境基本准备好，再次运行报错
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-lab/Test|$ ./vina.sh
'Deleting non-standard residues:AACE0_AACT1299_AI191300_ from pro
==============================
*** Open Babel Error in OpenAndSetFormat
Cannot open testset.sdf
0 molecules converted
/home/zhihao/Downloads/slurm-lab/Test/mols
0 molecules converted
ls: cannot access '*.mol2': No such file or directory
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-lab/Test|$重命名 testset200.sdf 为 testset.sdf 后 vina.sh 可运行
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-lab/Test|$ ./vina.sh
'Deleting non-standard residues:AACE0_AACT1299_AI191300_ from pro
201 molecules converted
/home/zhihao/Downloads/slurm-lab/Test/mols
201 molecules converted
201 files output. The first is vs1.mol2
#################################################################
# If you used AutoDock Vina in your work, please cite: #
# #
# O. Trott, A. J. Olson, #
# AutoDock Vina: improving the speed and accuracy of docking #
# with a new scoring function, efficient optimization and #
# multithreading, Journal of Computational Chemistry 31 (2010) #
# 455-461 #
# #
# DOI 10.1002/jcc.21334 #
# #
# Please see http://vina.scripps.edu for more information. #
#################################################################
Reading input ... done.
Setting up the scoring function ... done.
Analyzing the binding site ... done.
Using random seed: 1931149362
Performing search ...
0% 10 20 30 40 50 60 70 80 90 100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
done.
Refining results ... done.
mode | affinity | dist from best mode
| (kcal/mol) | rmsd l.b.| rmsd u.b.
-----+------------+----------+----------
1 -9.1 0.000 0.000
2 -8.1 3.635 8.475
3 -8.0 5.082 7.895
4 -8.0 5.733 8.730
5 -8.0 8.292 11.400
6 -8.0 3.347 7.427
7 -7.9 5.242 9.106
8 -7.9 5.712 8.642
9 -7.8 3.966 5.124
10 -7.7 5.592 8.513
Writing output ... done.
10 molecules converted
#################################################################
# If you used AutoDock Vina in your work, please cite: #
# #
# O. Trott, A. J. Olson, #
# AutoDock Vina: improving the speed and accuracy of docking #
# with a new scoring function, efficient optimization and #
# multithreading, Journal of Computational Chemistry 31 (2010) #
# 455-461 #
# #
# DOI 10.1002/jcc.21334 #
# #
# Please see http://vina.scripps.edu for more information. #
#################################################################
Reading input ... done.
Setting up the scoring function ... done.
Analyzing the binding site ... done.
Using random seed: -29891784
Performing search ...
0% 10 20 30 40 50 60 70 80 90 100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
done.
Refining results ... done.
mode | affinity | dist from best mode
| (kcal/mol) | rmsd l.b.| rmsd u.b.
-----+------------+----------+----------
1 -10.9 0.000 0.000
2 -10.9 0.026 1.961
3 -10.2 5.412 8.221
4 -10.2 5.414 8.148
5 -9.9 1.571 3.055
6 -9.8 4.895 8.583
7 -9.7 5.319 8.305
8 -9.6 4.863 6.782
9 -9.4 1.866 8.184
10 -8.9 4.055 5.777
Writing output ... done.
10 molecules converted
#################################################################
# If you used AutoDock Vina in your work, please cite: #
# #
# O. Trott, A. J. Olson, #
# AutoDock Vina: improving the speed and accuracy of docking #
# with a new scoring function, efficient optimization and #
# multithreading, Journal of Computational Chemistry 31 (2010) #
# 455-461 #
# #
# DOI 10.1002/jcc.21334 #
# #
# Please see http://vina.scripps.edu for more information. #
#################################################################
Reading input ... done.
Setting up the scoring function ... done.
Analyzing the binding site ... done.
Using random seed: -1740369596
Performing search ...
0% 10 20 30 40 50 60 70 80 90 100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
done.
Refining results ... done.
mode | affinity | dist from best mode
| (kcal/mol) | rmsd l.b.| rmsd u.b.
-----+------------+----------+----------
1 -7.7 0.000 0.000
2 -7.3 2.941 8.248
3 -7.3 2.000 7.731
4 -7.2 2.511 5.326
5 -7.0 2.891 5.561
6 -7.0 2.057 7.131
7 -6.8 3.090 8.148
8 -6.8 4.068 7.826
9 -6.7 1.553 2.107
10 -6.7 2.871 6.717
Writing output ... done.
10 molecules converted
#################################################################
# If you used AutoDock Vina in your work, please cite: #
# #
# O. Trott, A. J. Olson, #
# AutoDock Vina: improving the speed and accuracy of docking #
# with a new scoring function, efficient optimization and #
# multithreading, Journal of Computational Chemistry 31 (2010) #
# 455-461 #
# #
# DOI 10.1002/jcc.21334 #
# #
# Please see http://vina.scripps.edu for more information. #
#################################################################
Reading input ... done.
Setting up the scoring function ... done.
Analyzing the binding site ... done.
Using random seed: -832291788
Performing search ...
0% 10 20 30 40 50 60 70 80 90 100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
done.
Refining results ... done.
mode | affinity | dist from best mode
| (kcal/mol) | rmsd l.b.| rmsd u.b.
-----+------------+----------+----------
1 -7.8 0.000 0.000
2 -7.5 2.556 5.568
3 -7.2 3.261 7.045
4 -7.1 1.552 3.516
5 -7.1 3.531 8.237
6 -7.1 3.034 7.416
7 -7.1 3.005 5.912
8 -7.1 2.290 5.425
9 -7.0 2.334 7.098
10 -6.8 2.803 5.541
Writing output ... done.
10 molecules converted
#################################################################
# If you used AutoDock Vina in your work, please cite: #
# #
# O. Trott, A. J. Olson, #
# AutoDock Vina: improving the speed and accuracy of docking #
# with a new scoring function, efficient optimization and #
# multithreading, Journal of Computational Chemistry 31 (2010) #
# 455-461 #
# #
# DOI 10.1002/jcc.21334 #
# #
# Please see http://vina.scripps.edu for more information. #
#################################################################
Reading input ... done.
Setting up the scoring function ... done.
Analyzing the binding site ... done.
Using random seed: -1005170334
Performing search ...
0% 10 20 30 40 50 60 70 80 90 100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
done.
Refining results ... done.
mode | affinity | dist from best mode
| (kcal/mol) | rmsd l.b.| rmsd u.b.
-----+------------+----------+----------
1 -8.4 0.000 0.000
2 -8.4 8.694 13.384
3 -8.3 2.793 4.213
4 -8.0 8.455 13.245
5 -7.9 2.876 12.058
6 -7.7 2.990 12.238
7 -7.7 9.146 13.795
8 -7.7 10.056 14.591
9 -7.5 8.024 12.610
10 -7.5 7.944 9.897
Writing output ... done.
10 molecules converted
#################################################################
# If you used AutoDock Vina in your work, please cite: #
# #
# O. Trott, A. J. Olson, #
# AutoDock Vina: improving the speed and accuracy of docking #
# with a new scoring function, efficient optimization and #
# multithreading, Journal of Computational Chemistry 31 (2010) #
# 455-461 #
# #
# DOI 10.1002/jcc.21334 #
# #
# Please see http://vina.scripps.edu for more information. #
#################################################################
Reading input ... done.
Setting up the scoring function ... done.
Analyzing the binding site ... done.
Using random seed: 1161529493
Performing search ...
0% 10 20 30 40 50 60 70 80 90 100%
|----|----|----|----|----|----|----|----|----|----|
*******************????????观察进程负载，vina 仅使用一个核
zhihao@slurm-controller|/home/zhihao/Downloads/autodock_vina_1_1_2_linux_x86/bin|$ pidstat 1
Linux 4.18.0-553.el8_10.x86_64 (slurm-controller) 04/27/2025 _x86_64_ (4 CPU)
10:30:57 AM UID PID %usr %system %guest %wait %CPU CPU Command
10:30:58 AM 1000 245427 91.18 0.00 0.00 0.00 91.18 3 vina
10:30:58 AM 1000 245523 0.00 0.98 0.00 0.00 0.98 1 pidstat
10:30:58 AM UID PID %usr %system %guest %wait %CPU CPU Command
10:30:59 AM 1000 245427 92.08 0.00 0.00 0.00 92.08 3 vina
10:30:59 AM UID PID %usr %system %guest %wait %CPU CPU Command
10:31:00 AM 1000 245427 91.00 0.00 0.00 0.00 91.00 3 vina
10:31:00 AM 1000 245523 0.00 1.00 0.00 0.00 1.00 1 pidstat
^C
Average: UID PID %usr %system %guest %wait %CPU CPU Command
Average: 1000 245427 91.42 0.00 0.00 0.00 91.42 - vina
Average: 1000 245523 0.00 0.66 0.00 0.00 0.66 - pidstat
zhihao@slurm-controller|/home/zhihao/Downloads/autodock_vina_1_1_2_linux_x86/bin|$
修改 vina --cpu 参数为 3，重新运行 vina.sh
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-lab/Test|$ ./vina.sh
...
Writing output ... done.
10 molecules converted
#################################################################
# If you used AutoDock Vina in your work, please cite: #
# #
# O. Trott, A. J. Olson, #
# AutoDock Vina: improving the speed and accuracy of docking #
# with a new scoring function, efficient optimization and #
# multithreading, Journal of Computational Chemistry 31 (2010) #
# 455-461 #
# #
# DOI 10.1002/jcc.21334 #
# #
# Please see http://vina.scripps.edu for more information. #
#################################################################
Reading input ... done.
Setting up the scoring function ... done.
Analyzing the binding site ... done.
Using random seed: 1922223658
Performing search ...
0% 10 20 30 40 50 60 70 80 90 100%
|----|----|----|----|----|----|----|----|----|----|
***************************************************
done.
Refining results ... done.
mode | affinity | dist from best mode
| (kcal/mol) | rmsd l.b.| rmsd u.b.
-----+------------+----------+----------
1 -10.3 0.000 0.000
2 -10.2 5.951 9.637
3 -9.8 3.881 9.708
4 -9.6 2.802 3.936
5 -9.5 5.175 7.909
6 -9.5 6.486 10.253
7 -9.5 3.979 9.961
8 -9.4 4.542 8.731
9 -9.4 4.472 6.863
10 -9.4 3.443 5.373
Writing output ... done.
==============================
*** Open Babel Warning in PerceiveBondOrders
Failed to kekulize aromatic bonds in OBMol::PerceiveBondOrders (title is out.pdbqt)
==============================
*** Open Babel Warning in PerceiveBondOrders
Failed to kekulize aromatic bonds in OBMol::PerceiveBondOrders (title is out.pdbqt)
==============================
*** Open Babel Warning in PerceiveBondOrders
Failed to kekulize aromatic bonds in OBMol::PerceiveBondOrders (title is out.pdbqt)
==============================
*** Open Babel Warning in PerceiveBondOrders
Failed to kekulize aromatic bonds in OBMol::PerceiveBondOrders (title is out.pdbqt)
==============================
*** Open Babel Warning in PerceiveBondOrders
Failed to kekulize aromatic bonds in OBMol::PerceiveBondOrders (title is out.pdbqt)
==============================
*** Open Babel Warning in PerceiveBondOrders
Failed to kekulize aromatic bonds in OBMol::PerceiveBondOrders (title is out.pdbqt)
==============================
*** Open Babel Warning in PerceiveBondOrders
Failed to kekulize aromatic bonds in OBMol::PerceiveBondOrders (title is out.pdbqt)
==============================
*** Open Babel Warning in PerceiveBondOrders
Failed to kekulize aromatic bonds in OBMol::PerceiveBondOrders (title is out.pdbqt)
==============================
*** Open Babel Warning in PerceiveBondOrders
Failed to kekulize aromatic bonds in OBMol::PerceiveBondOrders (title is out.pdbqt)
==============================
*** Open Babel Warning in PerceiveBondOrders
Failed to kekulize aromatic bonds in OBMol::PerceiveBondOrders (title is out.pdbqt)
10 molecules converted
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-lab/Test|$output.sdf 输出看起来正常
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-lab/Test/mols|$ grep -c out.pdbqt output.sdf
1990
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-lab/Test/mols|$ ls *.mol2 | wc -l
201
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-lab/Test|$ ls mols/
output.sdf vs126.mol2 vs152.mol2 vs179.mol2 vs22.mol2 vs49.mol2 vs75.mol2
vs100.mol2 vs127.mol2 vs153.mol2 vs17.mol2 vs23.mol2 vs4.mol2 vs76.mol2
vs101.mol2 vs128.mol2 vs154.mol2 vs180.mol2 vs24.mol2 vs50.mol2 vs77.mol2
vs102.mol2 vs129.mol2 vs155.mol2 vs181.mol2 vs25.mol2 vs51.mol2 vs78.mol2
vs103.mol2 vs12.mol2 vs156.mol2 vs182.mol2 vs26.mol2 vs52.mol2 vs79.mol2
vs104.mol2 vs130.mol2 vs157.mol2 vs183.mol2 vs27.mol2 vs53.mol2 vs7.mol2
vs105.mol2 vs131.mol2 vs158.mol2 vs184.mol2 vs28.mol2 vs54.mol2 vs80.mol2
vs106.mol2 vs132.mol2 vs159.mol2 vs185.mol2 vs29.mol2 vs55.mol2 vs81.mol2
vs107.mol2 vs133.mol2 vs15.mol2 vs186.mol2 vs2.mol2 vs56.mol2 vs82.mol2
vs108.mol2 vs134.mol2 vs160.mol2 vs187.mol2 vs30.mol2 vs57.mol2 vs83.mol2
vs109.mol2 vs135.mol2 vs161.mol2 vs188.mol2 vs31.mol2 vs58.mol2 vs84.mol2
vs10.mol2 vs136.mol2 vs162.mol2 vs189.mol2 vs32.mol2 vs59.mol2 vs85.mol2
vs110.mol2 vs137.mol2 vs163.mol2 vs18.mol2 vs33.mol2 vs5.mol2 vs86.mol2
vs111.mol2 vs138.mol2 vs164.mol2 vs190.mol2 vs34.mol2 vs60.mol2 vs87.mol2
vs112.mol2 vs139.mol2 vs165.mol2 vs191.mol2 vs35.mol2 vs61.mol2 vs88.mol2
vs113.mol2 vs13.mol2 vs166.mol2 vs192.mol2 vs36.mol2 vs62.mol2 vs89.mol2
vs114.mol2 vs140.mol2 vs167.mol2 vs193.mol2 vs37.mol2 vs63.mol2 vs8.mol2
vs115.mol2 vs141.mol2 vs168.mol2 vs194.mol2 vs38.mol2 vs64.mol2 vs90.mol2
vs116.mol2 vs142.mol2 vs169.mol2 vs195.mol2 vs39.mol2 vs65.mol2 vs91.mol2
vs117.mol2 vs143.mol2 vs16.mol2 vs196.mol2 vs3.mol2 vs66.mol2 vs92.mol2
vs118.mol2 vs144.mol2 vs170.mol2 vs197.mol2 vs40.mol2 vs67.mol2 vs93.mol2
vs119.mol2 vs145.mol2 vs171.mol2 vs198.mol2 vs41.mol2 vs68.mol2 vs94.mol2
vs11.mol2 vs146.mol2 vs172.mol2 vs199.mol2 vs42.mol2 vs69.mol2 vs95.mol2
vs120.mol2 vs147.mol2 vs173.mol2 vs19.mol2 vs43.mol2 vs6.mol2 vs96.mol2
vs121.mol2 vs148.mol2 vs174.mol2 vs1.mol2 vs44.mol2 vs70.mol2 vs97.mol2
vs122.mol2 vs149.mol2 vs175.mol2 vs200.mol2 vs45.mol2 vs71.mol2 vs98.mol2
vs123.mol2 vs14.mol2 vs176.mol2 vs201.mol2 vs46.mol2 vs72.mol2 vs99.mol2
vs124.mol2 vs150.mol2 vs177.mol2 vs20.mol2 vs47.mol2 vs73.mol2 vs9.mol2
vs125.mol2 vs151.mol2 vs178.mol2 vs21.mol2 vs48.mol2 vs74.mol2
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-lab/Test|$ ls
2w17.pdb mols receptor.pdbqt testset.sdf vina-slurm.sh
conf.txt pro.pdb testset.mol2 vina.sh
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-lab/Test|$
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-lab/Test/mols|$ stat output.sdf
File: output.sdf
Size: 7159169 Blocks: 13984 IO Block: 4096 regular file
Device: fd00h/64768d Inode: 38064062 Links: 1
Access: (0664/-rw-rw-r--) Uid: ( 1000/ zhihao) Gid: ( 1000/ zhihao)
Context: unconfined_u:object_r:user_home_t:s0
Access: 2025-04-27 20:59:31.902990351 -0400
Modify: 2025-04-27 13:12:11.517281587 -0400
Change: 2025-04-27 13:12:11.517281587 -0400
Birth: 2025-04-27 10:59:01.029139138 -0400
zhihao@slurm-controller|/home/zhihao/Downloads/slurm-lab/Test/mols|$vina.sh 脚本并行化
分析知 for 循环内过程为分子匹配过程，相互独立，可以改为并行逻辑
移动 vina 项目到 /mnt/slurm_shared/ 共享文件夹
zhihao@slurm-controller|/home/zhihao/Downloads|$ rsync -auv slurm-lab /mnt/slurm_shared/
sending incremental file list
slurm-lab/
slurm-lab/Test/
slurm-lab/Test/2w17.pdb
slurm-lab/Test/conf.txt
slurm-lab/Test/pro.pdb
slurm-lab/Test/receptor.pdbqt
slurm-lab/Test/testset.sdf
slurm-lab/Test/vina-slurm.sh
slurm-lab/Test/vina.sh
sent 1,881,539 bytes received 161 bytes 3,763,400.00 bytes/sec
total size is 1,880,537 speedup is 1.00
zhihao@slurm-controller|/home/zhihao/Downloads|$
zhihao@slurm-controller|/home/zhihao/Downloads|$ mv autodock_vina_1_1_2_linux_x86 mgltools_x86_64Linux2_1.5.6 /mnt/slurm_shared/
zhihao@slurm-controller|/home/zhihao/Downloads|$vina.sh 改写为三个脚本，vina-start.sh, vina-mol.sh, vina-post.sh，其中 vina-mol.sh 可并行化。
vina-start.sh
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test|$ cat vina-start.sh
#!/bin/bash
#SBATCH --job-name=vina_docking
#SBATCH --output=logs/vina_start_%j.out
#SBATCH --error=logs/vina_start_%j.err
set -e
export PATH=/mnt/slurm_shared/autodock_vina_1_1_2_linux_x86/bin:$PATH
export PATH=/mnt/slurm_shared/mgltools_x86_64Linux2_1.5.6//bin:$PATH
export PATH=/mnt/slurm_shared/mgltools_x86_64Linux2_1.5.6//MGLToolsPckgs/AutoDockTools/Utilities24:$PATH
if [ -d "mols" ]; then
rm -rf mols
fi
mkdir mols
prepare_receptor4.py -r pro.pdb -e -o receptor.pdbqt ########## prepare only once
obabel -isd testset.sdf -omol2 -O testset.mol2
cd mols
obabel ../testset.mol2 -O vs.mol2 -m
jobid=$(sbatch --parsable /mnt/slurm_shared/slurm-lab/Test/vina-mol.sh )
sbatch --dependency=afterok:$jobid ../vina-post.sh
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test|$vina-mol.sh
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test|$ cat vina-mol.sh
#!/bin/bash
#SBATCH --array=1-201
#SBATCH --output=../logs/vina_mol_%j.out
#SBATCH --error=../logs/vina_mol_%j.err
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=1
#SBATCH --ntasks-per-node=1
echo "Slurm Array Task ID: $SLURM_ARRAY_TASK_ID"
export PATH=/mnt/slurm_shared/autodock_vina_1_1_2_linux_x86/bin:$PATH
export PATH=/mnt/slurm_shared/mgltools_x86_64Linux2_1.5.6//bin:$PATH
export PATH=/mnt/slurm_shared/mgltools_x86_64Linux2_1.5.6//MGLToolsPckgs/AutoDockTools/Utilities24:$PATH
task_id=$SLURM_ARRAY_TASK_ID
mol2_num=$(ls vs*.mol2 | wc -l)
echo "now in $(pwd)"
if [[ $task_id -gt 0 && $task_id -le $mol2_num ]]; then
prepare_ligand4.py -l vs${task_id}.mol2 -o ligand${task_id}.pdbqt
srun vina --config ../conf.txt --cpu 1 --ligand ligand${task_id}.pdbqt --out out${task_id}.pdbqt --log log.txt
obabel -ipdbqt out${task_id}.pdbqt -osd -O temp${task_id}.sdf
rm -f log.txt ligand${task_id}.pdbqt out${task_id}.pdbqt
else
echo "SLURM_ARRAY_TASK_ID ($SLURM_ARRAY_TASK_ID) is out of range."
fi
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test|$vina-post.sh
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test|$ cat vina-post.sh
#!/bin/bash
echo "now in $(pwd)"
cat temp*.sdf > output.sdf
cd ..
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test|$sbatch 提交 vina-start.sh，双节点八核都在工作。 检查 tempxx.sdf 正常生成，结束后会产生 output.sdf 结果文件，测试结束。
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test|$ sbatch vina-start.sh
Submitted batch job 11419
[root@slurm-controller slurm-cluster]# squeue
JOBID PARTITION NAME USER ST TIME NODES NODELIST(REASON)
11420_[86-201] debug vina-mol zhihao PD 0:00 1 (Resources)
11421 debug vina-pos zhihao PD 0:00 1 (Dependency)
11420_85 debug vina-mol zhihao R 0:01 1 slurm-compute2
11420_84 debug vina-mol zhihao R 0:35 1 slurm-compute1
11420_83 debug vina-mol zhihao R 0:38 1 slurm-compute1
11420_82 debug vina-mol zhihao R 0:40 1 slurm-compute1
11420_81 debug vina-mol zhihao R 0:46 1 slurm-compute1
11420_79 debug vina-mol zhihao R 1:57 1 slurm-compute2
11420_78 debug vina-mol zhihao R 2:09 1 slurm-compute2
11420_77 debug vina-mol zhihao R 2:23 1 slurm-compute2
[root@slurm-controller slurm-cluster]#
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test/mols|$ ls temp*
temp10.sdf temp22.sdf temp34.sdf temp46.sdf temp58.sdf temp6.sdf temp81.sdf
temp11.sdf temp23.sdf temp35.sdf temp47.sdf temp59.sdf temp70.sdf temp85.sdf
temp12.sdf temp24.sdf temp36.sdf temp48.sdf temp5.sdf temp71.sdf temp86.sdf
temp13.sdf temp25.sdf temp37.sdf temp49.sdf temp60.sdf temp72.sdf temp87.sdf
temp14.sdf temp26.sdf temp38.sdf temp4.sdf temp61.sdf temp73.sdf temp88.sdf
temp15.sdf temp27.sdf temp39.sdf temp50.sdf temp62.sdf temp74.sdf temp89.sdf
temp16.sdf temp28.sdf temp3.sdf temp51.sdf temp63.sdf temp75.sdf temp8.sdf
temp17.sdf temp29.sdf temp40.sdf temp52.sdf temp64.sdf temp76.sdf temp90.sdf
temp18.sdf temp2.sdf temp41.sdf temp53.sdf temp65.sdf temp77.sdf temp91.sdf
temp19.sdf temp30.sdf temp42.sdf temp54.sdf temp66.sdf temp78.sdf temp92.sdf
temp1.sdf temp31.sdf temp43.sdf temp55.sdf temp67.sdf temp79.sdf temp9.sdf
temp20.sdf temp32.sdf temp44.sdf temp56.sdf temp68.sdf temp7.sdf
temp21.sdf temp33.sdf temp45.sdf temp57.sdf temp69.sdf temp80.sdf
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test/mols|$
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test/mols|$ head temp7.sdf
out7.pdbqt
OpenBabel04282512433D
28 29 0 0 1 0 0 0 0 0999 V2000
4.0900 27.5220 6.3970 C 0 0 2 0 0 3 0 0 0 0 0 0
5.1170 26.8390 7.3330 C 0 0 2 0 0 3 0 0 0 0 0 0
5.5320 25.4610 6.7550 C 0 0 2 0 0 3 0 0 0 0 0 0
4.6310 27.6260 5.1030 O 0 0 0 0 0 0 0 0 0 0 0 0
5.9580 25.5860 5.2770 C 0 0 1 0 0 3 0 0 0 0 0 0
4.8790 26.3600 4.4770 C 0 0 2 0 0 3 0 0 0 0 0 0
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test/mols|$output.sdf 已生成
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test/mols|$ du -sh output.sdf
6.9M output.sdf
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test/mols|$ grep pdbqt output.sdf | uniq -c | wc -l
199
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test/mols|$ head output.sdf
out100.pdbqt
OpenBabel04282513053D
31 32 0 0 0 0 0 0 0 0999 V2000
1.5130 23.3700 12.3830 C 0 0 0 0 0 2 0 0 0 0 0 0
1.0140 24.2610 11.2350 C 0 0 0 0 0 0 0 0 0 0 0 0
0.4640 25.4400 11.5890 N 0 0 0 0 0 0 0 0 0 0 0 0
0.3830 25.7230 12.5560 H 0 0 0 0 0 0 0 0 0 0 0 0
1.1200 23.8700 10.0690 O 0 0 0 0 0 0 0 0 0 0 0 0
-0.0180 26.3120 10.6830 N 0 0 0 0 0 2 0 0 0 0 0 0
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test/mols|$千万分子对接架构思考
将大分子文件拆分为较小的分子文件，存储到分布式存储，slurm 计算集群主节点运行 slurm 控制器和数据库，计算节点运行 slurmd 服务，配置自动伸缩组动态扩展资源。计算结果写入分布式存储。
问题排查
srun 运行超时
排查后为 firewalld 导致，分析思路为：
tcpdump 看 slurm-controller 节点拒绝 35683 端口连接
[root@slurm-controller log]# tcpdump -n -i enp0s8 not port 22 -c 40 |& less
...
05:21:59.758261 IP 192.168.1.101.50952 > 192.168.1.100.35683: Flags [S], seq 352325565
6, win 29200, options [mss 1460,sackOK,TS val 2648148599 ecr 0,nop,wscale 7], length 0
05:21:59.758282 IP 192.168.1.100 > 192.168.1.101: ICMP host 192.168.1.100 unreachable
- admin prohibited filter, length 68端口 35683 为 srun 进程
[root@slurm-controller slurm-cluster]# netstat -nap | grep srun
tcp 0 0 0.0.0.0:38217 0.0.0.0:* LISTEN 202777/srun
tcp 0 0 0.0.0.0:33805 0.0.0.0:* LISTEN 202777/srun
tcp 0 0 0.0.0.0:34135 0.0.0.0:* LISTEN 202777/srun
tcp 0 0 0.0.0.0:45913 0.0.0.0:* LISTEN 202777/srun
tcp 0 0 0.0.0.0:35683 0.0.0.0:* LISTEN 202777/srun
[root@slurm-controller slurm-cluster]#firewalld-cmd 启用日志
[root@slurm-controller log]# firewall-cmd --set-log-denied=all
Warning: ALREADY_SET: all
success
[root@slurm-controller log]# sudo firewall-cmd --get-active-zones
public
interfaces: enp0s8 enp0s3
[root@slurm-controller log]#dmesg -Tw 验证端口被 firewalld drop
[Sun Apr 27 05:21:54 2025] FINAL_REJECT: IN=enp0s8 OUT= MAC=08:00:27:d5:e1:eb:08:00:27:22:8a:d0:08:00 SRC=192.168.1.101 DST=192.168.1.100 LEN=60 TOS=0x00 PREC=0x00 TTL=64 ID=44477 DF PROTO=TCP SPT=50952 DPT=35683 WINDOW=29200 RES=0x00 SYN URGP=0
[Sun Apr 27 05:21:55 2025] FINAL_REJECT: IN=enp0s8 OUT= MAC=08:00:27:d5:e1:eb:08:00:27:22:8a:d0:08:00 SRC=192.168.1.101 DST=192.168.1.100 LEN=60 TOS=0x00 PREC=0x00 TTL=64 ID=62684 DF PROTO=TCP SPT=50968 DPT=35683 WINDOW=29200 RES=0x00 SYN URGP=0ansible 内允许32768-60999/tcp 高端口通过后解决
srun -o 参数位置问题
-o 参数位置写在执行文件后面会导致没有重定向输出且无报错
复现
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test|$ sbatch /mnt/slurm_shared/test.sh -o ./alog
Submitted batch job 11140
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test|$ echo $?
0
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test|$ ls ./alog
ls: cannot access './alog': No such file or directory
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test|$ sbatch -o ./alog /mnt/slurm_shared/test.sh
Submitted batch job 11142
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test|$ ls ./alog
./alog
zhihao@slurm-controller|/mnt/slurm_shared/slurm-lab/Test|$因为 -o 参数被认为是执行文件的参数了，逻辑上也合理，注意即可
sbatch 输出重定向问题
问题为 sbatch 提交脚本后无法运行且无任何输出，squeue 内任务显示状态PD，NODELIST 为 none
无输出导致排查困难，只能注释文本内容测试。排查后发现因为脚本内标准输入输出重定向路径写错，logs 文件夹不存在，创建文件夹后正常运行。
#SBATCH --output=logs/vina_mol_%j.out
#SBATCH --error=logs/vina_mol_%j.err